Minimum Viable Description

Minimum viable description (MVD) is an emerging framework for audio description that establishes the foundational level of visual information needed to provide equal access to video content without introducing bias or cognitive overload. Rather than attempting to describe everything visible on screen, MVD focuses on conveying the most essential visual elements — such as scene changes, on-screen text, and key actions — that a viewer needs to understand the narrative. MVD is particularly relevant to automated audio description systems, where it guides decisions about what information to deliver automatically versus what to make available through on-demand interactions like visual question answering.